Welcome to Teachable Machine Lite Package

By: Meqdad Darwish

Teachable Machine Lite Package Logo

Downloads MIT License PyPI

Description

Run your Google Teachable Machine image models on Raspberry Pi, robots and other edge devices in a few lines of Python. Lightweight, powered by LiteRT, no TensorFlow required.

Teachable Machine Lite lets you take an image model you trained in Google's Teachable Machine, with no coding needed, and run it on a Raspberry Pi, a robot, or any small computer with just a few lines of Python. It runs on LiteRT (formerly TensorFlow Lite) instead of the full TensorFlow, so it installs quickly, starts fast and works comfortably on resource-constrained hardware. Pass it a camera frame from OpenCV or Picamera2 and get back the predicted class and its confidence, ready to draw on the image or act on. That makes it a good fit for robotics, IoT and other edge AI projects. It's the lightweight sibling of the Teachable Machine package.

Source Code is published on GitHub

Table Of Contents

  1. How-To Guide
  2. Requirements
  3. Code Examples
  4. Explanation
  5. Changelog
  6. Contributing
  7. Releasing

Supported Classifiers

Image Classification: Use an exported TensorFlow Lite model from Teachable Machine Platform (a model file with tflite extension). Both the Quantized and the Floating point exports are supported.

Works on Raspberry Pi OS (64-bit recommended), Linux, Windows and macOS, with NumPy 1.x or 2.x.